CN103020939A - Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data - Google Patents
Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data Download PDFInfo
- Publication number
- CN103020939A CN103020939A CN201210551692XA CN201210551692A CN103020939A CN 103020939 A CN103020939 A CN 103020939A CN 201210551692X A CN201210551692X A CN 201210551692XA CN 201210551692 A CN201210551692 A CN 201210551692A CN 103020939 A CN103020939 A CN 103020939A
- Authority
- CN
- China
- Prior art keywords
- image
- spissatus
- data
- sub
- cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a method for removing large-area thick clouds for optical remote sensing images through multi-temporal data. If large-area clouds exist in optical remote sensing images, and non-cloud data exist in other multi-temporal images in the areas, cloud area data can be repaired and reconstructed through complementary information of the data. The method comprises using all temporal non-cloud data for dictionary learning, taking relevance among images into account adaptively, learning an over-complete dictionary and optimal sparse representation coefficients of the images, and repairing and reconstructing the data in thick cloud areas. According to the method for removing large-area thick clouds, the different complementary information in multi-temporal image thick-cloud areas is used, relevance of the images serves as the weight, the image thick-cloud area data are filled by aid of a novel sparse representation theory, and accordingly, not only the high accuracy is obtained, but also the idea for removing large-area thick clouds is expanded, and an important practical significance is provided.
Description
Technical field
The present invention proposes a kind of multidate data of utilizing and removes the spissatus method of optical remote sensing image large tracts of land, the correlativity in phase data and the spissatus district of pending image when effectively utilizing other with the dictionary learning method of self-adaptation spectrum weighting, realize filling up of spissatus district data, relate to the remote sensing image processing technology field.
Background technology
When actual imaging, optical remote sensing image is subject to the impact of weather condition usually, and wherein the most common influence factor is exactly cloud.The cloud that the remote sensing image that most of optical sensors obtain all can be subject in various degree pollutes, and cloud layer has been covered the real information of lower floor's atural object.Especially spissatus, covered the atural object radiation information fully, cause the disappearance of image information, obviously reduced the quality of image, had a strong impact on follow-up image analysing computer work and further application.Therefore, effectively remove spissatusly, for improving the quality of image, effectively extract the terrain object object information, avoid the data waste, improve availability and the utilization factor of image, have great importance.
At present, the spissatus removal method of remote sensing image mainly contains two large classes in the world:
Spissatus for the single width image owing to there is not other reference information, painted angle to image self from benefit often and removes.The method that mainly contains interpolation of development comparative maturity, such as be close to most according to spissatus district neighborhood cloud-free area data, golden interpolation (Kriging) etc. in bilinearity, bicubic interpolation and the gram; The method of partial differential equation (PDE) by determining direction and the information of spissatus district adjacent diffusion, realizes the automatically inwardly diffusion of spissatus district border peripheral information, reaches the purpose that the cloud sector is filled up; Total variation (TV) method progressively inwardly carry out anisotropic diffusion by spissatus district to be repaired frontier point, but boundary is often fuzzyyer; The method that texture is synthetic can keep structure and the texture information in spissatus district preferably.These class methods have been utilized spissatus district and on every side the statistical properties, spatial entities distribution character and the autocorrelation of cloud-free area, and there is reasonable removal effect in less spissatus district, but for large-area spissatus weary and unable.
And for the multidate image data, people then consider to utilize the complementary information of this series image to carry out spissatus Transformatin.On the one hand, for pending spissatus zone in the image, can adopt same zone, the image data of phase substitutes when close, namely finishes the data reparation in spissatus district.Image of phase was more close in time when the method required other, and atural object is without significant change, and the spissatus zone of several images can not be overlapped, this so that its actual application value have a greatly reduced quality.Even found in addition the complementary image of satisfactory multidate, when replacing spissatus district, there is tone difference between the image of phase when close, obvious vestige can appear in the image of inlaying, and is difficult to the effect that reaches seamless spliced; Although can regulate by certain technology the effect of visualization of splicing, can have influence on further quantitative test and use.On the other hand, also have and utilize the multidate visual fusion to reach interpolation to go spissatus purpose.These class methods require the spissatus zone can not be overlapping equally, and spissatus removal is not thorough, affects easily the interpretation of image.
Fritter is spissatus to have preferably effect to said method for removing, but spissatus for the large tracts of land of common existence, often can not effectively utilize the complementary information between the multidate image, so can not thoroughly remove.
Summary of the invention
For existing shortcoming of removing cloud method, the present invention proposes a kind of multidate data of utilizing and removes the spissatus method of optical remote sensing image large tracts of land, fully excavate the complementary information of multidate image, adopt the adaptive weighted dictionary learning method of spectrum to fill the spissatus zone of large tracts of land of image.
Technical scheme of the present invention is that a kind of multidate data of utilizing are removed the spissatus method of optical remote sensing image large tracts of land, may further comprise the steps:
Step 1 is carried out geometry correction with the multidate image sequence of required processing, obtains the not image of phase simultaneously of the same area;
Step 2, to the same area not simultaneously the image of phase carry out spissatus district and detect, the cloud mask of phase image when obtaining each, and the related coefficient in the non-spissatus district of phase image when calculating each;
Step 3, with the same area not simultaneously the image of phase form the spatial spectral image of multidimensional according to time series, the spatial spectral image is divided the image sub-block, all image sub-blocks are reassembled as the matrix of two dimension; With step 2 gained each the time phase image the cloud mask form spatial spectral cloud mask according to time series, spatial spectral cloud mask is divided the mask sub-block, all mask sub-blocks are reassembled as the matrix of two dimension;
Step 4 is carried out dictionary learning, the spissatus zone of reconstructed image according to the matrix of above image sub-block and mask sub-block.
And, when step 2 is carried out the detection of spissatus district, the exceptional value in the image also is labeled as spissatus district.
And, in the step 3, adopt the window sliding that is of a size of n * n to be divided into size to the spatial spectral image and be the image sub-block of n * n * k, wherein, k is the Spectral dimension of image, be the not number of the image of phase simultaneously of the same area, then each image sub-block be reassembled as a column vector, and all column vectors sequentially formed the matrix of a two dimension by slip; And, adopt the window sliding that is of a size of n * n to be divided into size to spatial spectral cloud mask and be the mask sub-block of n * n * k then each mask sub-block to be reassembled as a column vector, and all column vectors are sequentially formed the matrix of a two dimension by slip.
And the size of establishing dictionary in the step 4 is kn
2* m, m 〉=256 and m>kn
2According to step 2 gained each the time phase image non-spissatus district related coefficient, determine adaptively the weight of multidate data in the dictionary learning process.
The multidate data of utilizing that the present invention proposes are removed the spissatus method of optical remote sensing image large tracts of land, and it is for areal, one group of time series image obtaining of phase does not carry out the method for dictionary learning, sparse reconstruction cloud Transformatin simultaneously.The image data of phase was more contiguous in time when the method needn't require difference, also needn't require the spissatus zone can not be overlapping, if several not simultaneously the image of phase have enough complementary informations, just can carry out reconstruction to spissatus area data preferably.And in the process of spissatus district reconstruction, the multidate image is carried out the spectrum restructuring, thinking according to many/Hyperspectral imaging is processed it, weigh its spectral correlation with the related coefficient between image, adaptively according to not simultaneously the correlativity of phase image consider the size of its role in reparation, fill up when mutually spissatus district data when utilizing the complete dictionary of mistake of sparse expression theoretical foundation study finally to realize each.The multidate data of utilizing that the present invention proposes are removed the spissatus method of optical remote sensing image large tracts of land, can effectively utilize the complementary information of multidate image, realize the spissatus thorough removal of large tracts of land, obviously improve the radiation quality of image, significantly promote the integrated application potentiality of image, have important using value at aspects such as target identification, terrain classification, mineral prospecting, military surveillances.Therefore, the spissatus removal method of optical remote sensing image large tracts of land not only has very important learning value but also has important practical significance.
Description of drawings
Fig. 1 is the embodiments of the invention process flow diagrams.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing, the spissatus removal method of multidate optical remote sensing image large tracts of land is according to an embodiment of the invention further described.Should be appreciated that specific embodiment described herein only is used for explaining the present invention, be not intended to limit the present invention.
Technical solution of the present invention can adopt computer software technology to realize automatic operational scheme.Below in conjunction with Fig. 1 in detail embodiment is described in detail and removes the spissatus concrete steps of large tracts of land.
Step 1 is carried out geometry correction with the multidate image sequence of required processing, obtains the not image of phase simultaneously of the same area.
The multidate data of utilizing that the present invention proposes are removed the spissatus method of optical remote sensing image large tracts of land, before to image processing, need to carry out geometry correction, in order to obtain the data of multidate image the same area, concrete geometry correction is prior art.The same regional multidate image of embodiment is through 7 width of cloth of 400 * 400 after the geometry correction MODIS image of phase simultaneously not.
Step 2, to the same area not simultaneously the image of phase carry out spissatus district and detect, obtain the cloud mask data, and the related coefficient in the non-spissatus district of phase image when calculating each.
Embodiment to the same area 7 width of cloth not simultaneously the MODIS image of phase carry out spissatus detection.Embodiment adopts the ratio R of visible light the 2nd passage (0.841-0.876 μ m) and the 1st passage (0.620-0.670 μ m)
2/ R
1To carry out spissatus detection, its principle is that the reflectivity of cloud on these two passages is very approaching as the cloud index, and water body and vegetation have larger difference at these two passages, when this ratio R
2/ R
1Namely think spissatusly between 0.9 and 1.1, but this method is not suitable for desert area.During implementation, those skilled in the art can be according to the characteristics of different sensors, search related data and by virtue of experience knowledge determine the cloud formula of index of other images and the scope of spissatus district cloud index.In addition, also can there be some exceptional values or invalid value in the image, can simultaneously this class value be labeled as spissatus district in the processing, so that the follow-up reconstruction that carries out in the lump.After spissatus detection, obtain the cloud mask of each width of cloth image, calculate accordingly the related coefficient (CC) in each non-spissatus district of width of cloth image, in order to weigh the correlativity of each image, determine adaptively its role in removing the cloud repair process.Wherein, the related coefficient of image is calculated by formula (1).
X wherein, y are two width of cloth images (being video conversion that the one digit number group is processed here), and N is the number of pixels of image, x
i, y
iBe image x, i pixel among the y, the value of i are 1,2 ... N, μ
x, μ
yIt is respectively its corresponding image average.
Step 3, with the same area not simultaneously the image of phase form the spatial spectral image of multidimensional according to time series, the spatial spectral image is divided the image sub-block, all image sub-blocks are reassembled as the matrix of two dimension; With step 2 gained each the time phase image the cloud mask form spatial spectral cloud mask according to time series, spatial spectral cloud mask is divided the mask sub-block, all mask sub-blocks are reassembled as the matrix of two dimension.
Among the embodiment, 7 width of cloth images are recombinated according to time series spectrum, form one new 400 * 400 * 7 spatial spectral image, can it be divided into 157609 4 * 4 * 7 image sub-block by the sliding type that moves from left to right, from top to bottom a pixel at every turn, then according to being listed as to the column vector that the image sub-block is launched into 112 * 1, at last all column vectors are formed one 112 * 157609 two-dimensional matrix with the sequencing of image slide block, this matrix that is made of image sub-block vector namely is that the back needs matrix to be processed.Simultaneously, the cloud mask is also similarly processed, finally obtained one 112 * 157609 mask matrix, be i.e. the matrix of mask sub-block vector formation.
Step 4 is carried out dictionary learning, the spissatus zone of reconstructed image according to the matrix of above image sub-block and mask sub-block.
On the basis of above-mentioned steps, can carry out dictionary learning.Embodiment adopts the mode of KSVD dictionary learning.The purpose of KSVD namely is on the basis of sample Y, learns out one and meets sparse standard T
0Dictionary D, obtain simultaneously the expression coefficient X of dictionary, as shown in Equation (2).Wherein, x
iThe i column element of expression X, || x
i||
0Represent 0 norm, i.e. x
iThe number of middle nonzero element.
The size of establishing dictionary in the step 4 is kn
2* m, m 〉=256 and m>kn
2According to step 2 gained each the time phase image non-spissatus district related coefficient, determine adaptively the weight of multidate data in the dictionary learning process.
Among the embodiment, the size of setting dictionary is 112 * 256, and initial value is for being listed as to normalized discrete cosine transform (DCT) base.The process of dictionary learning mainly comprises following two processes at present:
1. in the situation of given dictionary, find the solution the expression coefficient of dictionary.Find the solution the expression coefficient of dictionary, at present general is tracing algorithm, mainly contain match tracing (MP), orthogonal matching pursuit (OMP), regularization orthogonal matching pursuit (ROMP), compressed sensing tracking (CoSaMP), subspace tracking (SP), base tracking (BP) and tree type tracking (TMP) etc., the precision of integration algorithm, complexity and efficient, suggestion adopt orthogonal matching pursuit (OMP) to get final product.
2. represent in the situation of coefficient at known dictionary, upgrade dictionary.Upgrading dictionary is to be undertaken by atom (atom of each row of dictionary), namely only upgrades each time an atom.Upgrade k(k=1 such as formula (3), 2 ..., 112) and the individual atomic time, at first calculating does not contain the residual error E that this atomic time image represents
k
Wherein, Y has represented to take into account the two-dimensional matrix of all image sub-blocks compositions of the weights of multidate data in dictionary learning; d
jJ atom of expression dictionary, j=1,2 ... 112; x
jThe j column element of expression X.
When adding sparse constraint, namely cloud mask sub-block matrix Ω
kThe time, residual error becomes
It is carried out SVD according to formula (4) decompose k atom d after the renewal
kBe the first row of left unitary matrix U, the expression coefficient of its correspondence is the product of right unitary matrix V and first element of singular value matrix Λ.
The renewal dictionary carries out successively according to the sequencing of atom, when all atoms all travel through renewal once, namely finishes the once renewal of dictionary.If obtain better experimental result, then need to carry out the dictionary updating of more times number.After 1. 2. process finishes, just can obtain the rarefaction representation of spatial spectral image on the complete dictionary D of mistake of study that 7 width of cloth images form.At last, utilize the product of this dictionary and expression coefficient thereof just can rebuild spissatus district.
One of ordinary skill in the art will appreciate that, utilize the present invention not only harsh unlike conventional method to the requirement of multidate image data, it is spissatus reasonably to utilize correlativity between the multidate image to remove in the remote sensing image large tracts of land.
Large tracts of land in the optical remote sensing image is spissatus to have a widely versatility to the said method that the present invention proposes for removing, and is subjected to the less-restrictive of objective factor.Show that by the actual test result of simulated experiment the method has higher precision, the related coefficient between the image after the reparation and the intact image can reach more than 0.99.
Should be noted that and understand, in the situation that does not break away from the desired the spirit and scope of the present invention of accompanying claim, can make to the present invention of foregoing detailed description various modifications and improvement.Therefore, the scope of claimed technical scheme is not subjected to the restriction of given any specific exemplary teachings.
Claims (4)
1. one kind is utilized the multidate data to remove the spissatus method of optical remote sensing image large tracts of land, it is characterized in that, may further comprise the steps:
Step 1 is carried out geometry correction with the multidate image sequence of required processing, obtains the not image of phase simultaneously of the same area;
Step 2, to the same area not simultaneously the image of phase carry out spissatus district and detect, the cloud mask of phase image when obtaining each, and the related coefficient in the non-spissatus district of phase image when calculating each;
Step 3, with the same area not simultaneously the image of phase form the spatial spectral image of multidimensional according to time series, the spatial spectral image is divided the image sub-block, all image sub-blocks are reassembled as the matrix of two dimension; With step 2 gained each the time phase image the cloud mask form spatial spectral cloud mask according to time series, spatial spectral cloud mask is divided the mask sub-block, all mask sub-blocks are reassembled as the matrix of two dimension;
Step 4 is carried out dictionary learning, the spissatus zone of reconstructed image according to the matrix of above image sub-block and mask sub-block.
2. the described multidate data of utilizing are removed the spissatus method of optical remote sensing image large tracts of land according to claim 1, it is characterized in that: when step 2 is carried out the detection of spissatus district, the exceptional value in the image also is labeled as spissatus district.
3. the described multidate data of utilizing are removed the spissatus method of optical remote sensing image large tracts of land according to claim 2, it is characterized in that: in the step 3, adopt the window sliding that is of a size of n * n to be divided into size to the spatial spectral image and be the image sub-block of n * n * k, wherein, k is the Spectral dimension of image, be the not number of the image of phase simultaneously of the same area, then each image sub-block be reassembled as a column vector, and all column vectors sequentially formed the matrix of a two dimension by slip;
And, adopt the window sliding that is of a size of n * n to be divided into size to spatial spectral cloud mask and be the mask sub-block of n * n * k then each mask sub-block to be reassembled as a column vector, and all column vectors are sequentially formed the matrix of a two dimension by slip.
4. the described multidate data of utilizing are removed the spissatus method of optical remote sensing image large tracts of land according to claim 3, and it is characterized in that: the size of establishing dictionary in the step 4 is kn
2* m, m 〉=256 and m>kn
2According to step 2 gained each the time phase image non-spissatus district related coefficient, determine adaptively the weight of multidate data in the dictionary learning process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210551692.XA CN103020939B (en) | 2012-12-18 | 2012-12-18 | Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210551692.XA CN103020939B (en) | 2012-12-18 | 2012-12-18 | Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103020939A true CN103020939A (en) | 2013-04-03 |
CN103020939B CN103020939B (en) | 2015-06-10 |
Family
ID=47969506
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210551692.XA Active CN103020939B (en) | 2012-12-18 | 2012-12-18 | Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103020939B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103729831A (en) * | 2014-01-21 | 2014-04-16 | 中国科学院地理科学与资源研究所 | Thick cloud remote sensing removal method and device |
CN104156923A (en) * | 2014-08-12 | 2014-11-19 | 西北工业大学 | Multispectral remote sensing image cloud removing method based on sparse representation |
CN104616265A (en) * | 2015-02-12 | 2015-05-13 | 武汉大学 | Time domain reconstruction method of remote sensing sequence data |
CN104657952A (en) * | 2015-03-03 | 2015-05-27 | 武汉大学 | Remote-sensing daily snow product cloud removing method combined with temporal-spatial information |
CN104881850A (en) * | 2015-05-26 | 2015-09-02 | 上海大学 | Thick-cloud automatic removing method of multi-temporal remote sensing images |
CN106780353A (en) * | 2016-11-10 | 2017-05-31 | 哈尔滨工业大学 | A kind of multidate cloud block data reconstruction method based on phase spectrum angular metric |
CN106846285A (en) * | 2016-12-30 | 2017-06-13 | 苏州中科天启遥感科技有限公司 | High-performance remote sensing image synthetic method and device |
CN107230195A (en) * | 2017-07-12 | 2017-10-03 | 中国科学院遥感与数字地球研究所 | A kind of image treatment method and device |
CN108280810A (en) * | 2018-01-09 | 2018-07-13 | 北方工业大学 | Automatic processing method for repairing cloud coverage area of single-time phase optical remote sensing image |
CN109859118A (en) * | 2019-01-03 | 2019-06-07 | 武汉大学 | A kind of method and system for effectively inlaying polygon optimization removal cloud covered areas domain based on quaternary tree |
CN112381815A (en) * | 2020-11-26 | 2021-02-19 | 浙江易智信息技术有限公司 | Multi-dimensional fusion remote sensing big data cloud removing method |
CN112633123A (en) * | 2020-12-18 | 2021-04-09 | 国网浙江省电力有限公司电力科学研究院 | Heterogeneous remote sensing image change detection method and device based on deep learning |
CN114511786A (en) * | 2022-04-20 | 2022-05-17 | 中国石油大学(华东) | Remote sensing image cloud removing method integrating multi-temporal information and sub-channel dense convolution |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509262A (en) * | 2011-10-17 | 2012-06-20 | 中煤地航测遥感局有限公司 | Method for removing thin cloud of remote sensing image |
CN102799903A (en) * | 2012-07-18 | 2012-11-28 | 中国科学院自动化研究所 | High-spectrum automatic cloud detection method based on space and spectral information |
-
2012
- 2012-12-18 CN CN201210551692.XA patent/CN103020939B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509262A (en) * | 2011-10-17 | 2012-06-20 | 中煤地航测遥感局有限公司 | Method for removing thin cloud of remote sensing image |
CN102799903A (en) * | 2012-07-18 | 2012-11-28 | 中国科学院自动化研究所 | High-spectrum automatic cloud detection method based on space and spectral information |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103729831B (en) * | 2014-01-21 | 2017-04-05 | 中国科学院地理科学与资源研究所 | The spissatus minimizing technology of remote sensing and its device |
CN103729831A (en) * | 2014-01-21 | 2014-04-16 | 中国科学院地理科学与资源研究所 | Thick cloud remote sensing removal method and device |
CN104156923A (en) * | 2014-08-12 | 2014-11-19 | 西北工业大学 | Multispectral remote sensing image cloud removing method based on sparse representation |
CN104156923B (en) * | 2014-08-12 | 2017-01-11 | 西北工业大学 | Multispectral remote sensing image cloud removing method based on sparse representation |
CN104616265A (en) * | 2015-02-12 | 2015-05-13 | 武汉大学 | Time domain reconstruction method of remote sensing sequence data |
CN104616265B (en) * | 2015-02-12 | 2018-02-09 | 武汉大学 | A kind of time domain method for reconstructing of remote sensing sequence data |
CN104657952A (en) * | 2015-03-03 | 2015-05-27 | 武汉大学 | Remote-sensing daily snow product cloud removing method combined with temporal-spatial information |
CN104657952B (en) * | 2015-03-03 | 2017-04-12 | 武汉大学 | Remote-sensing daily snow product cloud removing method combined with temporal-spatial information |
CN104881850B (en) * | 2015-05-26 | 2017-12-22 | 上海大学 | A kind of spissatus automatic removal method of multi-temporal remote sensing image |
CN104881850A (en) * | 2015-05-26 | 2015-09-02 | 上海大学 | Thick-cloud automatic removing method of multi-temporal remote sensing images |
CN106780353B (en) * | 2016-11-10 | 2019-10-22 | 哈尔滨工业大学 | A kind of multidate cloud block data reconstruction method based on phase spectrum angular amount |
CN106780353A (en) * | 2016-11-10 | 2017-05-31 | 哈尔滨工业大学 | A kind of multidate cloud block data reconstruction method based on phase spectrum angular metric |
CN106846285A (en) * | 2016-12-30 | 2017-06-13 | 苏州中科天启遥感科技有限公司 | High-performance remote sensing image synthetic method and device |
CN107230195A (en) * | 2017-07-12 | 2017-10-03 | 中国科学院遥感与数字地球研究所 | A kind of image treatment method and device |
CN107230195B (en) * | 2017-07-12 | 2020-09-18 | 中国科学院遥感与数字地球研究所 | Image processing method and device |
CN108280810A (en) * | 2018-01-09 | 2018-07-13 | 北方工业大学 | Automatic processing method for repairing cloud coverage area of single-time phase optical remote sensing image |
CN109859118A (en) * | 2019-01-03 | 2019-06-07 | 武汉大学 | A kind of method and system for effectively inlaying polygon optimization removal cloud covered areas domain based on quaternary tree |
CN112381815A (en) * | 2020-11-26 | 2021-02-19 | 浙江易智信息技术有限公司 | Multi-dimensional fusion remote sensing big data cloud removing method |
CN112633123A (en) * | 2020-12-18 | 2021-04-09 | 国网浙江省电力有限公司电力科学研究院 | Heterogeneous remote sensing image change detection method and device based on deep learning |
CN112633123B (en) * | 2020-12-18 | 2024-06-04 | 国网浙江省电力有限公司电力科学研究院 | Heterogeneous remote sensing image change detection method and device based on deep learning |
CN114511786A (en) * | 2022-04-20 | 2022-05-17 | 中国石油大学(华东) | Remote sensing image cloud removing method integrating multi-temporal information and sub-channel dense convolution |
Also Published As
Publication number | Publication date |
---|---|
CN103020939B (en) | 2015-06-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103020939B (en) | Method for removing large-area thick clouds for optical remote sensing images through multi-temporal data | |
CN110136170B (en) | Remote sensing image building change detection method based on convolutional neural network | |
CN112288647B (en) | Remote sensing image cloud and shadow restoration method based on gating convolution | |
CN111986099B (en) | Tillage monitoring method and system based on convolutional neural network with residual error correction fused | |
CN108830796B (en) | Hyperspectral image super-resolution reconstruction method based on spectral-spatial combination and gradient domain loss | |
Wang et al. | Sub-pixel mapping of remote sensing images based on radial basis function interpolation | |
Mertens et al. | Sub-pixel mapping and sub-pixel sharpening using neural network predicted wavelet coefficients | |
CN107590515B (en) | Hyperspectral image classification method of self-encoder based on entropy rate superpixel segmentation | |
WO2021048607A1 (en) | Motion deblurring using neural network architectures | |
CN110501072B (en) | Reconstruction method of snapshot type spectral imaging system based on tensor low-rank constraint | |
Wang et al. | Indicator cokriging-based subpixel mapping without prior spatial structure information | |
CN103020912B (en) | The remote sensing image restored method of a kind of combination wave band cluster and sparse expression | |
Carbonneau et al. | UAV‐based training for fully fuzzy classification of Sentinel‐2 fluvial scenes | |
CN104123705B (en) | A kind of super-resolution rebuilding picture quality Contourlet territory evaluation methodology | |
Kang et al. | Robust object-based multipass InSAR deformation reconstruction | |
Li et al. | Mapping water bodies under cloud cover using remotely sensed optical images and a spatiotemporal dependence model | |
CN103810755A (en) | Method for reconstructing compressively sensed spectral image based on structural clustering sparse representation | |
CN112686830B (en) | Super-resolution method of single depth map based on image decomposition | |
CN112633140A (en) | Multi-spectral remote sensing image urban village multi-category building semantic segmentation method and system | |
CN112818920B (en) | Double-temporal hyperspectral image space spectrum joint change detection method | |
CN104867119A (en) | Structural lack image filling method based on low rank matrix reconstruction | |
CN104680487A (en) | Non-local image inpainting method based on low-rank matrix recovery | |
US11636649B2 (en) | Geospatial modeling system providing 3D geospatial model update based upon predictively registered image and related methods | |
CN117496347A (en) | Remote sensing image building extraction method, device and medium | |
CN108830130A (en) | A kind of polarization EO-1 hyperion low-altitude reconnaissance image typical target detection method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |